\begin{table}[htbp]\centering
\def\sym#1{\ifmmode^{#1}\else\(^{#1}\)\fi}
\caption{Panel B: Dyadic Dependent Variables}
\begin{tabular}{l*{2}{ccc}}
\toprule
                    &\multicolumn{3}{c}{(1)}               &\multicolumn{3}{c}{(2)}               \\
                    &\multicolumn{3}{c}{Full Sample}       &\multicolumn{3}{c}{Benchmark Sample}  \\
                    &        mean&          sd&       count&        mean&          sd&       count\\
\midrule
LUL                 &      0.0157&      0.1243&      147153&      0.0178&      0.1324&       52326\\
path1               &      0.0157&      0.1243&      147153&      0.0178&      0.1324&       52326\\
path2               &      0.0930&      0.2905&      147153&      0.1015&      0.3019&       52326\\
path3               &      0.3803&      0.4854&      147153&      0.4081&      0.4915&       52326\\
LUL1                &      0.0016&      0.0404&      147153&      0.0020&      0.0452&       52326\\
LUL2                &      0.0072&      0.0843&      147153&      0.0076&      0.0870&       52326\\
LUL3                &      0.0041&      0.0640&      147153&      0.0047&      0.0685&       52326\\
LUL4                &      0.0022&      0.0466&      147153&      0.0032&      0.0564&       52326\\
IG                  &      0.0167&      0.1283&      147153&      0.0188&      0.1359&       52326\\
\midrule
Observations        &      147153&            &            &       52326&            &            \\
\bottomrule
\multicolumn{7}{l}{\footnotesize Statistics in (1) are computed on the full sample of data available for each variable,}\\
\multicolumn{7}{l}{\footnotesize while statistics in (2) are computed on the benchmark sample.}\\
\end{tabular}
\end{table}
